Based on the premise that what people think influences their actions, self-talk strategies have been developed to direct and facilitate human performance. In this article, we present a meta-analytic review of the effects of self-talk interventions on task performance in sport and possible factors that may moderate the effectiveness of self-talk. A total of 32 studies yielding 62 effect sizes were included in the final meta-analytic pool. The analysis revealed a positive moderate effect size (ES = .48). The moderator analyses showed that self-talk interventions were more effective for tasks involving relatively fine, compared with relatively gross, motor demands, and for novel, compared with well-learned, tasks. Instructional self-talk was more effective for fine tasks than was motivational self-talk; moreover, instructional self-talk was more effective for fine tasks rather than gross tasks. Finally, interventions including self-talk training were more effective than those not including self-talk training. The results of this study establish the effectiveness of self-talk in sport, encourage the use of self-talk as a strategy to facilitate learning and enhance performance, and provide new research directions.